[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: SEUL: About friendlyness
> Sender: firstname.lastname@example.org
> Date: Mon, 14 Sep 1998 12:35:29 -0500
> From: Kevin Forge <email@example.com>
> Reply-To: firstname.lastname@example.org
> CC: email@example.com, firstname.lastname@example.org, email@example.com
> Content-Type: text/plain; charset=us-ascii
> X-UIDL: ec6c35314c61a292273959f943c5914f
> firstname.lastname@example.org wrote:
> > > > one. So, always must be a way for the user growing to an advanced
> > > > (traditional?) way to work.
> > > As long as there's an xterm window, a user could disable xdm or any other
> > > feature that's been installed for their ease.. and even if SEUL found some way
> > > of preventing that -- there' debian and slackware for the advanced and SuSE for
> > > the intermediate.
> > Debian and Slackware for the pedantic. Most of the big guns of kernel
> > devel don't use them. That includes Linus.
> Linus uses RedHat because he is lazy and Alan Cox dose because he works
> for them.
> Just to preempt any rebuttals :)
Alan was using RedHat well before they employed him. Top kernel
developpers left Slackware more than two years ago. While some can
have left RedHat either for ideological sympathy to Debian or due to
RedHat poor quality control in its two last distributions the fact is
telling "advanced" people use difficult distributions is a fallacy.
Real top guns are too busy doing real work like kernel hacking and
have no time to lose in things like hand customizing a raw X desktop
or editing config files in order to ensure a daemon is started at
One of the problems for making Linux user friendly is the
"hackeristic" aura. We have to dispel it and one of the things we
have to learn is lose respect to people who use difficult software for
the sake of it.
Tonight I will publish in Independence a quantified evaluation of the
benefits of kernel recompiling. They are close to nil if the kernel
was half decently compiled. In particular the fact of compiling for
Pentium instead of 386 improves performance of the C parts by less
than 2%. Yes two. And this is still smaller if we take the overall
performance of the whole kernel, and still smaller if you count for
time spent in user mode to get the overall system performance.
That is for the kernel compiling myth.
Jean Francois Martinez
Project Independence: Linux for the Masses